The immune system produces both antibody-mediated and cell-mediated responses. Each type of immune response is regulated by a type of lymphocyte, B cells (for antibody-mediated response) and T cells (for cell-mediated response). B cells initially recognize an antigen when the antigen binds to the IgM and IgD molecules on the B cell's surface. Each B cell clone recognizes only specific antigens due to the unique idiotype of that clone. Upon recognition of the antigen, B cells internalize and process the antigen for presentation via MHC class II molecules. B cells can thereby function as an antigen presenting cell (“APC”) for T cells. T cells bind to portions of foreign proteins (antigens) when portions of the protein associate with a major histocompatibility complex molecule (“MHC”), typically on an APC, in which the antigen is digested into fragments and presented on the surface of the APC bound to its MHC.
Several types of cancers have their origin in the circulatory system. Among the major types are: leukemias, a neoplasm of the bone marrow and blood; myelomas, a cancer of B cells; and lymphomas, a group of cancers that originate in the lymphatic system. Lymphomas can be further classified into several groups; one of these groups is the non-Hodgkin's lymphomas which, in turn, forms a diverse group of cancers. Three broad categories of these lymphomas are defined according to the International Working Formulation for tumor classification, low grade, intermediate grade and high grade, which differ in their curability and aggressiveness (Cheson, et al., “Report of an International Workshop to Standardize Response Criteria for Non-Hodgkin's Lymphomas,” J. Clin Oncol. 17(4):1244, 1999). Overall, these lymphomas collectively rank fifth in the United States in terms of cancer incidence and mortality, and approximately 50,000 new cases are diagnosed each year.
In a recent study which examined fifty-one case isolates of high-grade non-Hodgkins's lymphoma (NHL), forty-three were shown to be derived from B cells while eight were shown to be derived from T cells (Brown et al., Histopathology 14:621-27, 1989). Therefore, treatments directed specifically towards pathological B cells would be valuable in the treatment of non-Hodgkin's lymphomas and myelomas.
Initial attempts in the field to develop an immunology-based treatment directed at antigens uniquely produced by malignant B cells involved laboriously isolating and purifying idiotypic (Id) proteins directly from the pathological B cells. This purified protein was first used in model systems to treat the associated lymphoma. It was demonstrated that this active immunization against idiotypic determinants on isolated proteins could produce resistance to tumor growth in a mouse model system (Daley et al., J. Immunol. 120(5):1620-24, 1978; Sakato et al., Microbiol Immunol. 23(9):927-31, 1979). This phenomenon of resistance to tumor growth has been subsequently reproduced in a number of additional experimental tumor models (Stevenson et al., J. Immunol. 130(2):970-03, 1983; George et al., J. Immunol. 141(6):2168-74, 1988; Kwak, et al., Blood 76(11):2411-17, 1990).
Among the first attempts at bringing this idea and technology into the clinic was very labor intensive and utilized mouse monoclonal antibodies generated against proteins isolated from the patients' individual lymphomas following biopsy. Meeker and coworkers generated mouse monoclonal anti-idiotype antibodies for treatment of eleven patients after most had already undergone conventional lymphoma therapy (Meeker et al., Blood 65:1349-63, 1985). Positive results were obtained in roughly half the patients, with one case of apparent remission. In some of the patients, however, the lymphoma cells developed a resistance to the antibody via switching the class of cell surface-expressed antibodies (Meeker et al., N. Engl J. Med. 312:1658-65, 1985).
Another way a B cell lymphoma clone developed resistance to anti-idiotypic antibodies is via a somatic mutation in the CDR2 region (Cleary et al., Cell 44:97-106, 1986), thereby evading recognition. While this passive immunity approach for treatment has the advantage that it only requires isolation and purification of a relatively minor amount of idiotypic protein from a patient for raising an immune response in a mouse, the usefulness for treating lymphomas with monoclonal antibodies directed at idiotypes is limited. In the absence of a robust and convenient way to produce large quantities of idiotypic protein, however, this could prove to be the only practical way to exploit the abilities of the immune system to directly attack the idiotype of a B cell lymphoma.
Kwak et al. pursued a different approach and attempted the active immunization of patients using proteins purified from their own unique lymphomas in spite of the logistical requirement for isolating large quantities of idiotypic proteins (Kwak et al., N. Engl, J. Med. 327:1209-15, 1992). Patients who had minimal or no disease following chemotherapy were treated by vaccination with autologous idiotype proteins. In order to obtain sufficient quantities of idiotypic proteins for vaccination, lymphoma cells obtained by biopsy were fused with an established cell line to facilitate their growth in tissue culture, and the secreted idiotype proteins were purified via chromatography. Large scale application of this method of immunization is precluded due to the extreme labor requirements, technical barriers, and prohibitive costs. Additionally, concerns have recently been raised concerning the viral loads associated with protein production in mammalian cells.
In a following paper, Hsu et al. reported on the phase I/II of the above clinical trial utilizing vaccination of the idiotype conjugated to keyhole limpet hemocyanin (KLH) in the treatment of B-cell lymphoma (Hsu et al., Blood 89:3129-35, 1997). After standard chemotherapy, 41 patients with refractory non-Hodgkin's B-cell lymphoma were vaccinated with a tumor-specific idiotype. As per Kwak et al (1992), supra, the tumor-specific idiotype antigens were obtained by chromatographic purification of proteins produced by the patients' hybridomas. These proteins were therefore composed of the entire variable and constant regions of the patient's own immunoglobulin from the patients' lymphomas. The results showed that the generation of an anti-idiotype response correlated with improved clinical outcome. The duration of freedom from disease progression and overall survival of all patients mounting an anti-idiotype cellular immune response were significantly prolonged compared to those patients who did not mount an immune response. This study confirms that patients with B-cell lymphomas can be induced to make a specific immune response against tumor idiotype (Id) protein. Furthermore, the ability to generate an anti-idiotype immune response correlates with a more favorable clinical outcome. However, to treat each individual patient, lymphoma cells obtained by biopsy must be fused to established cell lines in order to allow the production of sufficient protein to vaccinate a typical patient. This process would-be difficult or impractical to use on a commercial scale.
More recently, Bendandi et al. demonstrated idiotypic, patient-specific vaccination-induced remissions in patients with follicular lymphoma (Bendandi et al., Nat. Med. 5:1171-77, 1999). Following standard chemotherapy, twenty patients demonstrating complete clinical remission were vaccinated using patient-specific idiotypic proteins accompanied by granulocyte-monocyte colony-stimulating factor (GM-CSF; see infra) Molecular analysis of the translocations characteristic of this lymphoma was conducted prior to chemotherapy, at clinical remission and following, vaccination therapy. Eight of eleven patients with detectable, translocations after chemotherapy-induced remission were found to undergo complete molecular remission following this vaccination. Tumor-specific cytotoxic CD8+ and CD4+ T cells were found in 19 of 20 patients. Tumor-specific antibodies were also detected but were not found to be required for remission. Again, this study used idiotypic proteins made up of the entire variable and constant region of the immunoglobulin found associated with the patient's lymphoma and produced by heterohybridoma fusion.
Therefore, directing an immune response to the idiotype of cells is a promising approach, but the above techniques are limited by the requirement of producing sufficient quantities of idiotypic proteins from each patient's lymphoma cells.
The concept of anti-idiotypic immunity against B cell tumors has also been used in the case of multiple myeloma. Results have been reported by Kwak and coworkers regarding its use in enhancing the specific efficacy of allogeneic marrow grafts by pre-immunizing the donor with myeloma IgG isolated from the patient (Kwak et al., Lancet 345 (8956):1016-20, 1995). Also, Massaia and coworkers vaccinated patients in remission following high-dose chemotherapy, followed by peripheral blood stem cell transplantation (Massaia et al., Blood 94:673-83, 1999).
Granulocyte-monocyte colony-stimulating factor (GM-CSF), used above in Bendandi et al.'s study, is a hematopoietic growth factor which stimulates proliferation and differentiation of hematopoietic progenitor cells. This cytokine also plays a role in shaping cellular immunity by augmenting T-cell proliferation (Santoli et al., J. Immunol. 141(2):519-26, 1988). increasing expression of adhesion molecules on granulocytes and monocytes (Young et al., J. Immunol. 145(2):607-15, 1990; Grabstein et al., Science 232(4749):506-08, 1986), and augmenting antigen presentation (Morrissey et al., J. Immunol. 139(4):1113-9, 1987; Heufler et al., J. Exp. Med. 167(2):700-05, 1988; Smith et al., J. Immunol. 144(5):1777-82, 1990).
Cell-based vaccines genetically engineered to produce GM-CSF have been shown to induce cellular immune responses capable of eliminating systemic lymphomas in preclinical models. This effect is mediated exclusively through activation of the cellular arm of the immune system (Levitsky et al., J. Immuno. 156(10): 3858-65, 1996). Similarly, low doses of free GM-CSF have been shown to enhance the protective anti-tumor immunity induced by idiotype protein-KLH immunization because of its ability to enhance immunity through an effect on the CD8 cells (Kwak et al., Proc. Natl. Acad. Sci. USA 93(20):10972-77, 1996. In one study, GM-CSF was shown to be the best immunomodulator to generate anti-tumor immunity among those tested in a model system (Dranoff, G., Proc. Natl. Acad. Sci. USA 90(8):3539-43, 1993.)
GM-CSF has also been used as a portion of a chimeric protein used to generate an immune response in model systems. Chen and Levy (Chen and Levy, J. Immunol. 154(7):3105-17, 1995; U.S. Pat. No. 6,099,846) studied the production of mouse monoclonal antibodies using a chimeric protein containing a portion of GM-CSF plus a portion of an antigen of interest, namely an idiotypic region obtained from a murine B-cell tumor, 38C13, both fused to portions of human immunoglobulin chains. Chen and coworkers have also studied fusion proteins where the GM-CSF moiety has been replaced by portions of IL-2 or IL-4 (Chen et al., J. Immunol. 153(10):4775-87, 1994). One explanation for the requirement of including the GM-CSF moiety (or interleukin moiety) was to augment the effect- of low levels of chimeric protein produced by the mammalian cell expression system. However, the use of purified GM-CSF co-administered with a chimeric protein to enhance the immune response of a vaccination has not been demonstrated.
With the advent of recombinant DNA technology, heavy and light chain cDNA molecules can now be cloned from hybridomas or from combinatorial libraries employing the polymerase chain reaction (PCR). This recombinant DNA technology allows researchers to manipulate the effector function or the binding function of a selected monoclonal antibody. In addition, combinatorial libraries of immunoglobulins can be generated by cloning a large number of VL and VH genes, randomly assorting them to create a library of different binding specificities, expressing them in E. coli, then screening the stochastic library for clones with the desired binding affinities (Huse et al., Science 246(4935):1275-81, 1989). Using this recombinant approach, human antibodies were cloned with high affinity and specificity for tetanus toxoid from a randomized combinatorial library expressed in E. coli (Mullinax et al., Proc. Natl. Acad. Sci. 87(20):8095-99, 1990). The immunoglobulin genes were cloned from activated B-cells into bacteriophage vectors using the polymerase chain reaction (PCR) with specific primers. The H and L chains were randomly combined and co-expressed in E. coli to comprise a library of 1 members. This combinatorial library was screened with 125I-tetanus toxoid and 0.2% of the clones displayed binding activity (Mullinax et al., supra). In addition, murine monoclonal antibodies have also been identified using a similar approach (Huse et al., supra; Caton et al., Proc. Natl. Acad. Sci. 87(16):6450-54, 1990). Winter and co-workers used a plasmid vector to clone immunoglobulin domains by the polymerase chain reaction for expression in bacteria (Orlandi et al., Proc. Natl. Acad. Sci. 86(10):3833-37, 1989).
Newly developed E. coli antibody cloning systems are very useful for the identification of genes encoding desired binding specificities. However, antibodies produced in E. coli are not generally useful for therapeutic applications. Typically, only the antibody antigen binding fragments, Fab or Fv, can be produced as secreted products in bacteria. In the rare instance when a whole chain tetrameric IgG has been produced in E. coli, the CH2 domains are not glycosylated. Nonglycosylated antibodies lack the cytolytic activities antibody-directed cellular cytotoxicity (ADCC) and complement activation that make passive immunotherapy so powerful. Mammalian expression systems produce glycosolated antibody and thus circumvent this limitation of the bacterial system. However, recent modifications in the CBER division of the FDA's “Points to Consider” clearly signal their concerns about viral loads associated with monoclonal antibodies produced in mammalian cells. Moreover, it is expected that any engineered antibody produced in a mammalian expression system will be quite expensive ($1500-$5000 per dose). Alternative expression systems that circumvent the difficulties encountered with current mammalian and bacterial systems are therefore highly desirable.
The baculovirus expression system is an attractive alternative to antibody production in E. coli and mammalian cells. The expression of recombinant proteins using the baculovirus system has been demonstrated in the past several years and has emerged as an excellent choice for high yield production (1-100 mg/L) of biologically active proteins in eukaryotic cells. The baculovirus/insect cell system also circumvents the solubility problems often encountered when recombinant proteins are overexpressed in prokaryotes. In addition, insect cells contain the eukaryotic post-translational modification machinery responsible for correct folding, disulfide formation, glycosylation, β-hydroxylation, fatty acid acylation, prenylation, phosphorylation and amidation not present in prokaryotes. The production of a functional, glycosylated monoclonal antibody recognizing human colorectal carcinoma cells from a baculovirus expression system has been recently demonstrated (Nesbit, J. Immunol. Methods 151:201-208, 1992). Additionally, expression of recombinant IgA has also been demonstrated in baculovirus cells, and this IgA was correctly assembled into heavy chain/light chain heterodimers, N-glycosylated, and secreted (Carayannopoulos et al., Proc. Natl. Acad. Sci. 91:8348-52, 1994, PCT Publication No. WO 98/30577, U.S. Pat. No. 6,063,905). However, the use of baculovirus to express a chimeric idiotypic protein for use as an immunotherapeutic agent to modify a B cell pathology such as B cell malignancies and autoimmune diseases has not been demonstrated.